STAT6730 Project
Exploring park access data from the Trust for Public Land
Polina Berezina, Shan Tang
12/14/2021
Introduction
The Trust for Public Land (TPL) [1] tracks green space availability across U.S. using the proprietary ParkScore index. This index measures how well cities are meeting their residents’ need for public parks based on several categories:
- access: percent of residents that live within a 10-minute walking distance, or 0.5 miles, to a public park.
- acreage: median size, percent of parkland within the city area.
- investment: sum of per capita statistic on public, non-profit spendings to manage parklands, volunteer hours.
- amenities: per capita statistics on the number of basketball hoops, dog parks, playgrounds, restrooms, recreation and senior centers, and splashpads.
- equity (2021 only): average of four metrics. Per capita ratio of nearby park space between neighborhoods of color and white, as well as low-income vs. high-income neighborhoods. Percentage of people of color living within a 10-minute walk to a park and a similar percentage for low-income households.
Each city is awarded up to 100 points based on each one of these categories, and total of the scores is normalized to form an overall ParkScore rating. The data is taken from the Tidy Tuesday project [2]. In this report, we would like conduct exploratory data analysis to explore yearly trends across different categories of ParkScore and how the TPL has changed their methodology. We also explore importance of certain park features to their rankings. Finally, we analyze the newest 2021 data published by TPL to uncover how addition of the equity category to the ParkScore has affected cities’ rankings.
Regional patterns
WOULD NOT SHOW UP TO SPEED UP TESTING. REMOVE “EVAL=FALSE” IN THE END
Data cleaning and imputation
Year as a factor
Removed % and $ in the data frame, convert variables from string (%) to double (0-1)
rename the column names
More than 30% of data is missing in 6 out of 28 variables, including restroom, splashground and park benches counts. We omit these variables from analysis.
Assuming missing at random for others: impute using mice package with PMM Hot Deck method
Data analysis
Distribution of variables
Variables over the years. Increasing trend in park size, park access, playground scores
Park Scores vs. Features. TPL has different max. scores or feature weight across years. 2020-21 scores are higher across park size, easier accessibility and investment in parks features.
Amenities: (number of) basketball hoops, dog parks, playgrounds. Overall increasing trend in score with better amenities. But plateaus
Top ranking cities across years
Exploring equity in 2021
TPL added 4 features that go into the new equity score : park space ratio and 10-minute walk for low-income and people of color.
Hard to directly compare statistics since they changed weighting methods.
## [,1] [,2] [,3]
## [1,] "1" "Minneapolis" "Washington"
## [2,] "2" "Washington, D.C." "St. Paul"
## [3,] "3" "St. Paul" "Minneapolis"
## [4,] "4" "Arlington, Virginia" "Arlington"
## [5,] "5" "Cincinnati" "Chicago"
## [6,] "6" "Portland" "San Francisco"
## [7,] "7" "Irvine" "Irvine"
## [8,] "8" "San Francisco" "Cincinnati"
## [9,] "9" "Boston" "Seattle"
## [10,] "10" "Chicago" "Portland"Conclusion
Pay attention to the different scores range/weight across the years
Parks tend to have greater score with larger park size, easier accessibility, higher investment and more facilities.
TPL affected rankings of a lot of cities by including an equity metric in